We believe in good software

because we make it

Continuous Defect Prediction

Cheaper work force is not the way we are winning our international customers. It never was. From the beginning, we put on the highest quality and newest technologies. We are also working on our procedures and software engineering techniques. One of them is a method to predict the errors on the continuous integration (CI) server. We called it Continuous Defect Prediction. It uses machine Learning (ML) to create prediction models using historical continuous integration data. The models are then used on the developer machine to warn the developer if the changes he is making to the project, are likely to introduce a problem on the build server. This novel technique together with an accompanying dataset will be presented at the 14th International Conference on Mining Software Repositories that will take place on 21 and 21 of May in Buenos Aires, Argentina. This conference is collocated with the largest scientific software engineering conference the 39th International Conference on Software Engineering (ICSE) and we will take part in this conference also. Meet us in Buenos!

People – our most valuable asset

Five years ago we paraphrased a popular buzzword “Think global, act local” and said that we “Think Global, We Act… Global”. The last 5 years proofed us right. Out new USA customers can confirm that. But we are not forgetting about to act locally.

People are the most valuable CODEFUSION asset. We are recruiting our employees in 100% from the local academic market. Every year we are able to attract the best graduates from Opole University of Technology or (less frequently) from Opole University. How w do this? We are lucky to operate from reasonably small city but with two academic institutions “producing” 200 graduates a year. Those not willing to be sucked by big corporate players into surrounding big cities like Wroclaw or Katowice come to us. How do they know about us? We are a regular guest at various events organized at the Opole University of Technology. We are always at the IT Academic Day or Job Fairs. We were active sponsor of the Opole Robots Fair up until this year, where they decided to move outside the academia (and cut us off from our target – students). Tomorrow we give an open lecture at Opole University of Technology. We will be talking about our toolbox. This time it will be about source control systems. Join us!

Microsoft R Server

Machine learning, data mining, big data, deep learning – those are only the four buzzwords making the headlines in software engineering these days. It is the world of big data. It seems that we have more data than we can handle. What can we do about it?

One of the rules of CODEFUSION is to be on the very edge of the technology. It is the same this time. We have very ambitious project involving machine learning going. Slowly we are becoming specialists in R programming. R is popular computer language and environment for statistical computation. Since we are working on enterprise level solution we quickly realized we need a server R solution. We are very glad that Microsoft is here ached of the pack. We first tried their cloud machine learning implementation (AzureML) and now we are turning towards on premise Microsoft R Server (MSR). If the name does not ring the bell not worry, it is a relive young member of Microsoft Server family. The more surprising the current version of MSR… 9.0. As it turns out the Microsoft R Server is nothing else than rebranded (and rewritten in Microsoft .NET Core) Revolution R Enterprise (with deploy) after acquiring their creator Revolution Analytics by Microsoft in January 2015. This fits into Microsoft strategy investing heavily into R language. We have Microsoft R Open (based on the open source R version and compatible with CRAN distributed version 3.3.2) Microsoft added the libraries for their SQL Server and renamed deploy to operationalization. Under this mouthful name hides a technology to turn every R script into a web service. Those web services are then available over REST API and can be consumed using Swagger for example in any C# .NET application. We are very glad to have all those features for our project. We will report on the news in this project here. So, stay tuned!

Merry Christmas

spSection – our part of spColumn

Almost a year passed since spColumn v5.00 was released. Now we are happy to inform that a new version of this civil engineering software we’ve co-developed is out. It is spColumn 5.5.

In this version, we’ve completely rewritten the module named spSection. It is a module that lets the engineer model the irregular shape columns. We’ve developed tooling for modelling columns and reinforcement. Tools include features like creating rectangles, circles, polygons, joining shapes or modyfing nodes. Also there are transformation features like rotation, translation and reflection. We’ve added tooling for entering coordinates from keyboard, validating the model and importing/exporting the model from/to dxf (a de facto standard among civil engineers).

spColumn v5.50 can be tested for free in a Trail version. This version is downloadable from the Structurepoint website. Give it a try!